Download Anna Frebel nucleosynthesis, stars + chemical evolution

Nucleosynthesis,
stellar abundances,
and chemical evolution
Anna Frebel
P-329
Guest lecture
“Stars and planets” class
by Dimitar Sasselov, Fall 2010
Ulitmate question
How did the solar abundances come about?
Stellar spectra
What sort of stars are we looking for?
unevolved, low-mass
stars; <1 Msun
to ensure long
lifetimes
=>unmixed, too,
to avoid surface
abundances
contamination
with nuclear
burning products
© B.J. Mochejska (APOD)
Three Observational
Steps to Find Metal-Poor Stars
1. Sample selection and visual
inspection:
Find appropriate candidates
(Ca scales with Fe!)
2. Follow-up spectroscopy
(medium resolution):
Derive estimate for [Fe/H]
from the Ca II K line
3. High-resolution
spectroscopy:
Detailed abundances analysis
Frebel et al. 2005b
“Look-back time”
spectroscopic comparison
Abundances are derived from
integrated absorption line strengths
[Fe/H] = log(NFe/NH)  log(NFe/NH)
*
equals 1/250,000th
of the solar Fe
abundance
important spectral absorption lines in
stars
• H lines 6562Å, 4860Å, 4340Å, 4101Å
• CH g-band @ 4313Å and others
• Li @ 6707Å
• Mg b lines @ ~5170Å
• Ca K line @ 3933Å
• Na D lines @ ~5880Å
• Eu @ 4129Å
• Sr @ 4077Å, 4215Å
• Ba @ 4554Å
Fe lines are
everywhere in the
spectrum -- always
easily accessible
Carbon & nitrogen
 Huge carbon
abundance ([C/Fe]= +3.7):
(=> not so carbon-poor...)
Synthetic spectrum: red lines
Carbon (CH) band
5,000
and 12,000 times
more carbon and
nitrogen exist than
iron!
 Huge nitrogen
abundance ([N/Fe]= +4.1):
(=> not so nitrogen-poor...)
Synthetic spectrum: red lines
Reminder:
Solar ratio [C,N/Fe] = 0
Nitrogen (NH) band
Frebel et al. 2008, ApJ subm.
HE 1327-2326
Ca II K line
Calcium often used as
proxy for the Fe
abundance!
(..and Fe for metallicity)
–5.4
Interstellar Ca
(Ca scales with Fe!)
Frebel et al. (2005), Nature
Mg b lines
Eu
Thorium II Line 4019Å
Abundance:
Synthetic spectrum (based on atomic data and model atmosphere)
to match observed spectrum
Synthetic spectrum that
includes NO thorium
Th
HE 1523-0901
Frebel et al. (2010), in prep.
‘Best fit’ synthetic spectrum
Uranium in HE 1523-0901
Synthetic spectrum that includes NO uranium
Synthetic spectrum with U abundance if it had NOT decayed
Frebel et al. (2007)
‘Best fit’ synthetic spectrum
How do we interpret stellar spectra?
need to know:
model atmosphere analysis techniques
knowledge about nucleosynthesis,
stellar evolution,
chemical evolution,
cosmological understanding of galaxy
formation
Model atmosphere analysis
techniques
Stellar parameters fully characterize a star:
effective temperature Teff
surface gravity log g
metallicity [Fe/H]
(microturbulent velocity vmic)
ATomic data
• every absorption line is an atomic transition
• determined by atomic physics parameters
• Vienna Atomic Line Database (VALD)
http://vald.astro.univie.ac.at/~vald/php/vald.php
• National Institute for Standards and Technology (NIST)
http://www.nist.gov/pml/data/asd.cfm
From [email protected] Mon Nov 1 10:03:10 2010
Date: Tue, 31 Aug 2010 23:56:00 +0200
From: [email protected]
Subject: Re:
============= job.012302 =============
# begin request
# extract all
# default configuration
# short format
#
# 4057.0, 4058.5
# end request
Damping parameters Lande
Elm Ion WL(A) Excit(eV) log(gf) Rad. Stark Waals factor References
'Ti 1', 4057.0060, 2.3340, -4.645, 7.735,-5.924,-7.491, 0.230,' 1 1 1 1
'Si 2', 4057.0090, 12.8390, -1.330, 0.000, 0.000, 0.000,99.000,' 2 2 2 2
'F 3', 4057.0630, 54.8200, -0.340, 0.000, 0.000, 0.000,99.000,' 3 3 3 3
'V 1', 4057.0650, 2.1220, -0.203, 8.158,-5.083,-7.799, 1.000,' 4 4 4 4
'Cr 1', 4057.1370, 4.4460, -1.424, 8.330,-5.330,-7.720, 1.160,' 5 5 5 5
'Co 1', 4057.1820, 0.2240, -3.249, 4.653,-6.374,-7.867, 1.230,' 6 6 6 6
Stronger line <=>
lower excit <=>
higher log gf
1
2
3
4
5
6
1
2
3
4
5
6
1'
2'
3'
4'
5'
6'
Definitions: log

Stellar ‘abundances’ are number density calculations with
respect to H and the solar value
On a scale where H is 12.0:
log (X)  log 10 N X /N H   12
for element X
This quantity is the output of all model atmospheres!
i.e. MOOG code (of Chris Sneden, publicly available) +
Kurucz models (=inhouse!)
definitions: [fe/h]
QuickTime™ and a
decompressor
are needed to see this picture.
where NFe and NH is the no. of iron and hydrogen atoms per unit of volume respectively.
QuickTime™ and a
decompressor
are needed to see this picture.

 
NO
NO
N Fe
N Fe 
 log 10 (
) star  log 10 (
) sun  log 10 (
) star  log 10 (
) sun 
N
N
N
N

 

H
H
H
H

A /H   B /H   A /B
for elements A and B
Solar
abundances
Photospheric (=‘stellar’ abundance)
•
•
•
•
•
Anders, Grevesse & Sauval ‘89
Grevesse & Sauval ‘98
Asplund, Grevesse &Sauval ‘05
Grevesse, Asplund & Sauval ‘07
Asplund, Grevesse, Sauval & Scott ‘09
•
•
reference element: H
calculation
Meteoritic (=‘star dust’ grain analysis)
•
•
Lodders 03
Lodders, Palme & Gail 09
•
•
reference element: Si
measurement
•
Volatile elements depleted, incl. the most abundant
elements: H, He, C, N, O, Ne cannot rely on meteorites
to determine the primordial Solar System abundances
for such elements
For each application, the most similarly
obtained solar abundances should be use
to minimize systematic uncertainties!
how to calculate
chemical abundances
• Need a spectrum => measure equivalent width of
absorption lines (=integrated line strength)
• Need atomic data (excit. potential+log gf values) => feed
both into “model atmosphere”
• Get: calculated abundance (number density) log  (X)
• Calculate [Fe/H] with solar abundances
•
•
•
•
•
Example:
log  (Mg)star = 5.96; log  (Fe)star = 5.50
log  (Mg)sun = 7.60; log  (Fe)sun = 7.50
[Mg/H] = log  (Mg)star - log  (Mg)sun = -1.64
[Mg/Fe] = [Mg/H] - [Fe/H] = -1.64 - (-2.0) = 0.36
How metal-poor?
classical example:
early universe: primordial gas
how metal-poor is the next-generation star?
canonical SN Fe yield: 0.1 Msun
available gas mass: 106 Msun
M tot 10 6 M sun
NH 

mH
mH
QuickTime™ and a
decompressor
are needed to see this picture.
N Fe 
M tot 0.1M sun

mFe
56m H

log (Fe) sun  log( N Fe /N H ) sun 12  7.50

 log( N Fe /N H ) sun  7.50 12  4.50
N Fe 0.1M sun
mH
107

 6

NH
56m H
10 M sun
56
107
 [Fe /H]  log(
)  (4.50)  4.2
56
classification scheme
Range
[Fe/H] ≥ +0.5
[Fe/H] = 0.0
[Fe/H] ≤ –1.0
[Fe/H] ≤ –2.0
[Fe/H] ≤ –3.0
[Fe/H] ≤ –4.0
[Fe/H] ≤ –5.0
[Fe/H] ≤ –6.0
Term
Acronym
#
Super metal-rich
SMR
some
Solar
—
a lot!
Metal-poor
MP
very many
Very metal-poor
VMP
many
Extremely metal-poor EMP
~100
Ultra metal-poor
UMP
1
Hyper metal-poor
HMP
2
Mega metal-poor
MMP
--
Extreme Pop II stars!
as suggested by Beers & Christlieb 2005
Halo Metallicity distribution function (MDF)
Previous ‘as
observed’, raw MDF
is not a realistic
presentation!
(but shows that we have
been doing a good job in
finding these stars..)
Non-zero tail!!!
Schoerck et al. 2008
The most metal-poor stars are
extremely rare but extremely important!
“Surface Pollution”
through accretion
Bondi-Hoyle-Lyttleton accretion:
dM/dt = 4 G2M2 / v3
V1
Accretion for 10 billions years?
stellar orbit
Galactic disk
V2
Three-component potential:
disk, spheroid, halo
(Johnston 1998)
blue, bluer, the bluest
Lower metallicity leads to decreased opacity
 stars are hotter than solar equivalents
 look bluer (bluer colors)
 needs to be taken into account!
for temperature
measurements,
abundance analyses,
stellar populations
studies
Nucleosynthesis
All elements heavier than Li, Be, B are
made during stellar evolution and
supernova explosions
Stellar nucleosynthesis
QuickTime™ and a
decompressor
are needed to see this picture.
most important reactions in stellar nucleosynthesis:
* Hydrogen burning:
- The proton-proton chain
All textbooks, wikipedia
- The CNO cycle
....
* Helium burning:
- The triple-alpha process
- The alpha process
* Burning of heavier elements:
Timmes+ ~95
- Carbon burning process
Woosely&Weaver 1995
- Neon burning process
Heger & Woosley 2008
- Oxygen burning process
- Silicon burning process
* Production of elements heavier than iron:
- Neutron capture:
- The R-process
Many details not
- The S-process
known, but good models
- Proton capture:
out there
- The Rp-process
- Photo-disintegration:
- The P-process
neutron-capture processes
_
-decay: n => p + e- v
e
QuickTime™ and a
decompressor
are needed to see this picture.
• s-process: neutron-capture longer than beta-decay timescale
• r-process: neutron-capture shorter than beta-decay timescale
slow n-cap process
• in 1-8Msun AGB stars; AGB stars
are major providers of C and sprocess elements in the universe
(through mass loss)
• produce s-rich companions: CH
stars, Ba stars, s-rich metal-poor
stars
good knowledge of sprocess theoretically;
QuickTime™ and a
decompressor
important for calculating
are needed to see this picture.
the solar r-process
component
The two neutron sources
in AGB stars
13C(a,n)16O
Needs 13C !
Major neutron source
13C-pocket
Primary source!
T8 = 0.9-1
Interpulse phase
(1- 0.4) 105 yr
Radiative conditions
Nn = 107 cm-3
lower mass AGBs
22Ne(a,n)25Mg
Abundant 22Ne
Minor neutron source
Neutron burst
Secondary source
T8 = 3 (low 22Ne efficiency)
Thermal pulse
6 yr
Convective conditions
Nn (peak) = 1010 cm-3
higher mass AGBs
the AGB engine
He, 12C, 22Ne, s-process elements: Zr, Ba, ...
At the
stellar
surface:
C>O, sprocess
enhance
ments
thermally pulsing
AGB stars
r-process
QuickTime™ and a
None decompressor
are needed to see this picture.
r-Process Enhanced Stars
(rapid neutron-capture process)
 Responsible for the production of the heaviest elements
 Most likely production site: SNe II => pre-enrichment
 Chemical “fingerprint” of previous nucleosynthesis event
(only “visible” in the oldest stars because of low metallicity)
 ~5% of metal-poor stars with
[Fe/H] <  2.5 (Barklem et al. 05)
 Only 15-20 stars known so
far with [Eu/Fe] > 1.0
Nucleo-chronometry: obtain
stellar ages from decaying Th, U and
stable r-process elements (e.g. Eu, Os)
SN
star
-- decay --
today
[Th and U can also be measured in the Sun, but the chemical evolution has
progressed too far; required are old, metal-poor stars from times when only very few
SNe had exploded in the universe]
Our Cosmic Lab
The r-Process Pattern
Very good
agreement
with scaled
solar rprocess
pattern
for Z>56
scaled solar
r-process pattern
decayed Th,U
HE 1523-0901
According to
metal-poor star
abundances,
the r-process
is universal!
Frebel et al. (2007)
Precision at work!
CS 22892-052
HD 115444
Scaled solar
r-process
element
pattern!!
BD +17 3248
CS 31082-001
HD 221170
HE 1523-0901
Cowan 2007, priv. comm
They all have the same abundance pattern,
particularly among heavy neutron-capture elements!
r-process must be a universal process!
chemical evolution
origin of the elements
abundances trends
chemical evolution
Zentrum fuer Astronomie und Astrophysik, TU Berlin
All the atoms (except H, He & Li)
were created in stars!
Pop III: zero-metallicity stars
Pop II: old halo stars
Pop I: young disk stars
We are made of stardust!
 Old stars contain fewer elements
(e.g. iron) than younger stars
How and when did
these early stars form?
e.g. HE 1327-2326
Big Bang
First star
exploding
First chemical
enrichment
Heger & Woosley 2008
Primordial
gas cloud
2nd generation stars
forming from
enriched material
Why important?
Metal-poor stars provide the only available diagnosis for zerometallicity Pop III nucleosynthesis and early chemical enrichment
Pre-enrichment by a “faint SN”
Iwamoto et al. 2005 Science 309 451
• “Faint” SN with mixing and
fallback: Post-explosion
abundance distribution
–Explains high C, N, O,
Mg
(Smaller mass cut for
HE1327-2326 to account for
high [Mg/Fe])
–Explain other metal-poor
stars with [Fe/H]<3.5
–Neutron-capture
elements not included
M=25M, Z=0, low E
Iwamoto et al. 2005 Science 309 451
… with some (new) upper limits
Abundance trends
[Mg/Fe]
Alpha-elements
[Si/Fe]
Alpha elements multiple of
He: (C,O), Ne, Mg, Si, S, Ar,
Ca, Ti (not pure)
[Ca/Fe]
Synthesis during stellar
evolution and a-capture in
supernova explosion of
massive stars (>8 M)
Fe and a-elements produced
in the explosions of massive
stars (SN type II)
Fe-rich ejecta
from the SN of
low-mass stars
(SN type Ia)
Aoki, Frebel et al. 2006, ApJ
What is so special About
the most Fe-poor stars?
hyper
Fe-poor
ultra
Fe-poor
extremely
Fe-poor
very
Fe-poor
hyper
Fe-poor
ultra
Fe-poor
extremely
Fe-poor
very
Fe-poor
A compilation of abundances of ~800 metal-poor stars with
The very different chemical signature of the hyper iron-poor stars
[Fe/H]~<-2.0
can be found at
is crucial for understanding the formation of the elements!
http://www.cfa.harvard.edu/~afrebel/abundances/abund.html
(published in Frebel ‘10, review article on metal-poor stars)
plots with abundance trends
https://www.cfa.harvard.edu/~afrebel/abundances/abund.html
• Li in HE 1300 depleted
in accordance with the
star’s evolutionary
status (subgiant)
• Majority of depletion
seems to be taking
place in the range
5500-5600 K
• Li depletion does not
significantly depend on
metallicity
Frebel et al. 06, ApJ submitted
Lithium
the bigger picture
using stars to study the hierarchical
assembly of galaxy formation
“near-field cosmology”
A long time ago...
2nd and later generations
of stars (<1 M)
First stars
(100 M)
Big Bang
today
first galaxies
today’s galaxies
Larson & Bromm 2001
Cosmic time (not to scale)
metallicity-luminosity relation
Ultra-faint dwarfs
Stellar
archaeology with
the most metal-poor
stars in MW satellite
galaxies
Classical dSphs
Ultra-faint dwarfs
Martin et al. (2007)
Metallicity distribution function of dSph
galaxies
More metal-poor stars in the
ultra-faints than in halo!?!
Ultra-faint dwarfs
MW halo stars
Classical dwarfs
Kirby et al. (2008)
(targets selected from
Simon & Geha 2007)
“classical” dSph
have no extremely
metal-poor stars?!?
(Helmi et al. 2006)
=> yes, they do!
What can we learn from the existing dwarf
galaxies?
Stellar archaeology: examine the chemical history in
search for their oldest population to learn about
- early chemical evolution in small systems
- chemical signatures that relate dwarf galaxies to MW
If surviving dwarfs are analogs of early MW building
blocks then we should find chemical evidence of it!
Stellar metallicities & abundances of
metalpoor stars in dwarf galaxies
should
agree with those found in the MW halo
Mg, Ca, Ti (a-elements)
No discrepancy of ultra-faint
dwarf galaxy stars with those of
MW halo (at low metallicities)!
•
Stars in ultra-faint dwarfs studied
by AF and colleagues (Ursa
MajorII, Coma Berenices, LeoIV)
(Frebel+2010, Simon+2010)
•
Stars in ultra-faint dwarfs studied
by others (Hercules, Bootes)
(Koch+2008, Norris+2009)
•
Stars in classical dSphs (Sculptor,
Carinae, Draco, Sextans, Ursa Minor,
Fornax, Leo)
halo stars
ultra-faints
dSphs stars
(Shetrone+2001,2003, Venn+2004, Sadakane+04,
Aoki+2009)
•
Halo stars
(e.g. Cayrel+2004, Barklem+2005, Aoki+2005,
Lai+2008 plus many others!) See Frebel (2010)
review for a complete list of abundances and
references.
ultra-faint dwarf galaxy abundances
Excellent
agreement with
the MW chemical
Comparison with
evolution
Cayrel+ 04 halo data
(black open circles)
Spread in some
elements (C, ncapture elements)
red squares:
Ursa Major II
blue dots:
Coma Berenices
black diamond:
MW halo giant
Frebel et al. (2010a)
Summary
•
•
•
•
•
•
•
•
•
spectroscopy
abundance analyses
stellar atmospheres
stellar evolution
nucleosynthesis
SN physics and explosions
nuclear+atomic physics
chemical evolution
(near-field) cosmology